Visualizing Genomics Data
http://rockefelleruniversity.github.io/RU_RNAseq/
##The Course
- The Course
- Importance of Visualising Genomics Data.
- Reminder of file types
- Reminder of data types
- Materials
- Visualising genomics data in R
- Plotting genome axis
- Plotting genome data
- Plotting genome annotation
- Plotting genome sequence
- Plotting genomic alignments
- Plotting from external databases
##Reminder of file types
In this session we will be dealing with a range of data types. For more information on file types you can revisit our material.
For more information on visualising genomics data in browsers you can visit our IGV course.
- IGV.
##Reminder of data types in Bioconductor
We will also encounter and make use of many data structures and data types which we have seen throughout our courses on HTS data. You can revisit this material to refresh on HTS data analysis in Bioconductor and R below.
- Bioconductor.
- Genomic Intervals
- Genomic Scores
- Sequences
- Gene Models
- Alignments.
- ChIP-seq.
- ATAC-seq.
- RNA-seq.
##Materials.
All material for this course can be found on github. * Visualising Genomics Data
Or can be downloaded as a zip archive from here.
* Download zip
##Materials. - Presentations, source code and practicals.
Once the zip file in unarchived. All presentations as HTML slides and pages, their R code and HTML practical sheets will be available in the directories underneath.
- viz_course/presentations/Slides/ Presentations as an HTML slide show.
- viz_course/presentations/exercises/ Some tasks/examples to work through.
##Materials. - Data for presentations, practicals.
All data to run code in the presentations and in the practicals is available in the zip archive. This includes coverage as bigWig files, aligned reads as BAM files and genomic intervals stored as BED files.
We also include some RData files containing precompiled results from querying database (in case of external server downtime).
All data can be found under the Data directory
Data/
##Set the Working directory
Before running any of the code in the practicals or slides we need to set the working directory to the folder we unarchived.
You may navigate to the unarchived VisualisingGenomicsData folder in the Rstudio menu
Session -> Set Working Directory -> Choose Directory
or in the console.
setwd("/PathToMyDownload/VisualizingGenomicsData/viz_course/presentations/Slides")
# e.g. setwd("~/Downloads/VisualizingGenomicsData/viz_course/presentations/Slides") Covered so far. (Making the axis)
We saw previously how we can create GenomeAxisTrack objects to represents scales and plot these using the plotTracks() functions.
Covered so far. (Adding data from GRanges/bigWigs)
We can create DataTrack objects from GRanges objects and plot these alongside our GenomeAxisTrack object using the plotTracks() functions.
library(rtracklayer)
allChromosomeCoverage <- import.bw("data/activatedReads.bw",
as="GRanges")
accDT <- DataTrack(allChromosomeCoverage,chomosome="chr17")
plotTracks(c(accDT,genomeAxis),
from=47504051,to=47600688,
chromosome="chr17",type="hist") 
Covered so far. (Adding gene models to plots)
We created a GeneRegionTrack containing information on our gene models from our TxDb object and added this alongside data and axis to provide some context.
library(TxDb.Mmusculus.UCSC.mm10.knownGene)
customFromTxDb <- GeneRegionTrack(TxDb.Mmusculus.UCSC.mm10.knownGene,
chromosome="chr17")
plotTracks(c(accDT,customFromTxDb,genomeAxis),
from=47504051,to=47600688,
chromosome="chr17",type="hist") 
##SequenceTracks
When displaying genomics data it can be important to illustrate the underlying sequence for the genome being viewed.
Gviz uses SequenceTrack objects to handle displaying sequencing information.
First we need to get some sequence information for our genome of interest to display. Here we will use one of the BSgenome packages specific for hg19 - BSgenome.Hsapiens.UCSC.hg19. This contains the full sequence for hg19 as found in UCSC
## 159138663-letter "DNAString" instance
## seq: NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN...NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN##SequenceTracks - From a BSgenome object
We can create a SequenceTrack object straight from this BSgenome object using the SequenceTrack() constructor.
We can then plot this SequenceTrack, as with all tracks, using the plotTracks() functions. Here we specify the from, to and chromosome parameters to select a region to display.
sTrack <- SequenceTrack(Hsapiens)
plotTracks(sTrack,from=134887024,to=134887074,
chromosome = "chr7",cex=2.5) 
##SequenceTracks - From a DNAstringset object
We can also specify a DNAstringset object which we have encountered in the Bioconductor courses.
dsSet <- DNAStringSet(Hsapiens[["chr7"]])
names(dsSet) <- "chr7"
sTrack <- SequenceTrack(dsSet)
plotTracks(sTrack,from=134887024,to=134887074,
chromosome = "chr7",cex=2.5)
##SequenceTracks - From a DNAstringset object
We can also create our custom SequenceTrack from a Fasta file.
Here we use an example containing only the sequence around the region we are looking at to save space. Since the sequence is only of the region of interest we need specify the sequence limits for the from and to arguments. With completer fasta files, from and to would be set as for other SequenceTrack examples.
sTrack <- SequenceTrack("data/chr7Short.fa")
plotTracks(sTrack,from=1,to=50,
chromosome = "chr7",cex=3) 
##SequenceTracks - Displaying complement sequence
As with IGV, the sequence can be displayed as its complement. This is performed here by setting the complement argument to the plotTracks() function to TRUE/T.
sTrack <- SequenceTrack("data/chr7Short.fa")
plotTracks(sTrack,from=1,to=50,
chromosome = "chr7",complement=T,cex=3) 
##SequenceTracks - Displaying strand information
We can also add 5’ to 3’ direction as we have for plotting GenomeAxisTrack objects using the add53 parameter. This allows for a method to illustrate the strand of the sequence being diplayed.
sTrack <- SequenceTrack("data/chr7Short.fa")
plotTracks(sTrack,from=1,to=50,
chromosome = "chr7",complement=F,
add53=T,cex=2.5) 
##SequenceTracks - Displaying strand information
Notice the 5’ and 3’ labels have swapped automatically when we have specified the complement sequence.
sTrack <- SequenceTrack("data/chr7Short.fa")
plotTracks(sTrack,from=1,to=50,
chromosome = "chr7",complement=T,
add53=T,cex=2.5) 
##SequenceTracks - Controlling base display size
We can control the size of bases with the cex parameter, as with the standard R plotting.
An interesting feature of this is that when plotted bases overlap, Gviz will provide a colour representation of bases instead of the bases’ characters.
##SequenceTracks - Controlling base display size
We can force the bases to be always a color representation instead of letter irrespective of cex by setting the noLetters parameter to TRUE.
##SequenceTracks - Controlling base display size
If we want to colour bases by own colour scheme, we can provide a named vector of our desired colours to the fontcolor parameter.
colForLetters <- c(A = "darkgrey", C = "red",
T = "darkgrey",G = "darkgrey")
plotTracks(sTrack,from=1,to=50,
chromosome = "chr7",cex=3,
fontcolor=colForLetters) 
##AlignmentsTrack.
So far we have displayed summarised genomics data using GRange objects or GRanges with associated metadata.
A prominent feature of Gviz is that it can work with genomic alignments, providing methods to generate graphical summaries on the fly.
Genomic alignments are stored in Gviz within the AlignmentsTrack object.
Here we can read genomic alignments in from a BAM file, see our file formats course material, by specifying its location.
## ReferenceAlignmentsTrack 'AlignmentsTrack'
## | genome: NA
## | active chromosome: chrNA
## | referenced file: data/small_Sorted_SRR568129.bam
## | mapping: id=id, cigar=cigar, mapq=mapq, flag=flag, isize=isize, groupid=groupid, status=status, md=md, seq=seq##AlignmentsTrack. Plotting Aligned Reads in Gviz
The AlignmentsTrack object can be plotted in the same manner as tracks using plotTracks() function.
Since the BAM file may contain information from all chromosomes we need to specify a chromsome to plot in the chromosome parameter and here we specify the from and to parameters too.
##AlignmentsTrack. Plotting Aligned Reads in Gviz
By default AlignmentTracks are rendered as both the reads themselves and the calculated coverage/signal depth from these reads.
Reads, as with AnnotationTrack objects, show the strand of the aligned read by the direction of the arrow.
##AlignmentsTrack. Plotting Aligned Reads in Gviz
The type of plot/plots produced can be controlled by the type argument as we have done for DataTrack objects.
The valid types of plots for AlignmentsTrack objects are “pileup”, “coverage” and “sashimi” (We’ve come across sashimi plots before).
The type “pileup” displays just the reads.
##AlignmentsTrack. Plotting Aligned Reads in Gviz
The type “coverage” displays just the coverage (depth of signal over genomic positions) calculated from the genomic alignments.
##AlignmentsTrack. Plotting Aligned Reads in Gviz
As we have seen the default display is a combination of “pileup” and “coverage”.
We can provide multiple type arguments to the plotTracks() function as a vector of valid types. The order in vector here does not affect the display order in panels.
##AlignmentsTrack. Sashimi plots
We have seen sashimi plots in IGV when reviewing RNA-seq data.
Sashimi plots display the strength of signal coming from reads spanning splice junctions and so can act to illustrate changes in exon usage between samples.
In IGV, we previous made use of the BodyMap data to show alternative splicing of an exon between heart and liver.
##AlignmentsTrack. Sashimi plots in Gviz
To recapitulate this plot, we have retrieved the subsection of BodyMap data as BAM files.
First we must create two AlignmentsTrack objects, one for each tissue’s BAM file of aligned reads.
In this case since we are working with paired-end reads we must specify this by setting the isPaired parameter to TRUE
heartReads <- AlignmentsTrack("data/heart.bodyMap.bam",
isPaired = TRUE)
liverReads <- AlignmentsTrack("data/liver.bodyMap.bam",
isPaired = TRUE)
liverReads ## ReferenceAlignmentsTrack 'AlignmentsTrack'
## | genome: NA
## | active chromosome: chrNA
## | referenced file: data/liver.bodyMap.bam
## | mapping: id=id, cigar=cigar, mapq=mapq, flag=flag, isize=isize, groupid=groupid, status=status, md=md, seq=seq##AlignmentsTrack. Sashimi plots in Gviz
As with DataTrack objects we can combine the AlignmentTracks as a vector for plotting with the plotTracks() function.
By default we will display the reads and calculated coverage. Here the paired reads and split reads are illustrated by thick and thin lines respectively.
##AlignmentsTrack. Sashimi plots in Gviz
To reproduce a plot similar to that in IGV we can simply include the “sashimi” type in the type parameter vector, here alongside “coverage”
plotTracks(c(heartReads,liverReads),
chromosome="chr12",
from=98986825,
to=98997877,
type=c("coverage","sashimi")) 
##AlignmentsTrack. Highlighting genomic alignment information.
The AlignmentTrack object allows for specific parameters controlling how reads are displayed to be passed to the plotTracks() function.
A few useful parameters are col.gaps and col.mates or lty.gap and lty.mates which will allow us to better distinguish between gapped alignments (split reads) and gaps between read pairs respectively.
plotTracks(c(liverReads),
chromosome="chr12",
from=98986825,to=98997877,
col.gap="Red",col.mate="Blue") 
##AlignmentsTrack. Highlighting genomic alignment information.
Similarly using lty.gap and lty.mate parameters.
Line width may also be controlled with lwd.gap and lwd.mate parameters continuing the similarities to Base R plotting.
##AlignmentsTrack. Highlighting mismatches to reference.
A common purpose in visualising alignment data in broswers is review information relating to mismatches to the genome which may be related to SNPs.
In order to highlight mismatches to the genome reference sequence we must first provide some information on the reference sequence.
One method for this is to attach sequence information to the AlignmentsTrack itself by providing a SequenceTrack object to referenceSequence parameter in the AlignmentsTrack() constructor. Here we can create the SequenceTrack object ffor mouse.
library(BSgenome.Mmusculus.UCSC.mm10)
sTrack <- SequenceTrack(BSgenome.Mmusculus.UCSC.mm10)
activatedReads <- AlignmentsTrack("data/activatedSNPread.bam",
isPaired = TRUE,
referenceSequence=sTrack) ##AlignmentsTrack. Highlighting mismatches to reference.
Now when we can replot the pileup of reads where mismatches in the reads are highlighted.
##AlignmentsTrack. Highlighting mismatches to reference.
We could also specify the SequenceTrack in the plotTracks() function as shown for the liver reads example here. Here we simply include the relevant SequenceTrack object as a track to be plotted alongside the BAM.
##Bringing in External data.
Gviz has functions to allow us to import data from external repositories and databases.
As in the IGV course, visualising genomics data in the context of additional genome information and external data held at these repositories provides a deeper insight into our own data.
In this course we will look at two main methods of querying external databases-
- The BiomartGeneRegionTrack object and constructor.
- The UcscTrack object and constructor
##Bringing in External data. Gene models through Biomart
The biomaRt Bioconductor package to programatically query various Biomarts.
Gviz allows us to both query Biomarts and automatically create a GeneRegionTrack using the BiomartGeneRegionTrack objects and BiomartGeneRegionTrack() constructor.
##Bringing in External data. Gene models through Biomart
Here we construct a simple BiomartGeneRegionTrack object using the parameters to define locations of interest - “chromsome”, “start”,“end”,“genome” as well as the Biomart to use, in this case Ensembl by setting the name parameter.
bgrTrack <- BiomartGeneRegionTrack(genome="hg19",
start=26591341,
end=27034958,
chromosome = "chr7",
name="ENSEMBL") ## Warning in .Seqinfo.mergexy(x, y): The 2 combined objects have no sequence levels in common. (Use
## suppressWarnings() to suppress this warning.)##Bringing in External data. Gene models through Biomart
We can then plot the BiomartGeneRegionTrack as we have previous GeneRegionTracks.
## Warning in .Seqinfo.mergexy(x, y): The 2 combined objects have no sequence levels in common. (Use
## suppressWarnings() to suppress this warning.)
##Bringing in External data. Gene models through Biomart
We can also specify filters in the BiomartGeneRegionTrack() constructor using the filter parameter.
Gviz uses the BiomaRt Bioconductor package to query the Biomarts so we can review availble options using BiomaRt’s useMart() function to list available data collections and listDatasets() function to review all available datasets within the ENSEMBL_MART_ENSEMBL mart.
library(biomaRt)
martList <- listMarts()
mart = useMart("ENSEMBL_MART_ENSEMBL")
dataList <- listDatasets(mart)
mart = useMart("ENSEMBL_MART_ENSEMBL", dataset="hsapiens_gene_ensembl")
filterList <- listFilters(mart)##Bringing in External data. Gene models through Biomart
Here we select only genes which have been annotated by both havana and ensembl (so called Golden Transcripts)
bgrTrack <- BiomartGeneRegionTrack(genome="hg19",
start=26591341,
end=27034958,
chromosome = "chr7",
name="ENSEMBL",
filter=list(source="ensembl_havana")) ##Bringing in External data. Gene models through Biomart
Once we have retrieved our filtered gene models we can plot them as before.
##Bringing in External data. Tracks from UCSC
A well known browser and source of genomic data and annotation is the UCSC genome browser. Gviz can create track directly from UCSC tables using the functionality from rtracklayer Bioconductor package.
The Ucsctrack() constructor and object allow for the query and track construction of a variety of data types. The Ucsctrack() function therefore requires us to specify the track type we expect using the trackType parameter as well as the required UCSC table using the track parameter.
##Bringing in External data. Tracks from UCSC
To understand which tables are available we can query the rtracklayer package to identify track and table names.
We can use the broswerSession() function to connect to UCSC and the genome() function to establish the genome of interest.
We can then review available tracks using the trackNames() function.
##Bringing in External data. Tracks from UCSC
We can identify available tables by using the ucscTableQuery() function and the tableNames() function.
query <- ucscTableQuery(session, "Ensembl Genes",
GRangesForUCSCGenome("hg19", "chr7",
IRanges(26591341,27034958)))
tableNames(query) ## [1] "ensGene" "ccdsInfo" "ensGtp" "ensPep" "ensemblSource" "ensemblToGeneName"
## [7] "knownToEnsembl"##Bringing in External data. Tracks from UCSC
ucscTrack <- UcscTrack(genome = "hg19",
chromosome = "chr7",
track = "ensGene",
from = 26591341,
to = 27034958,
trackType = "GeneRegionTrack",
rstarts = "exonStarts",
rends = "exonEnds",
gene ="name",
symbol = "name2",
transcript = "name",
strand = "strand"
) To build the UCSC annotation as a GeneRegionTrack we must specify some information specific to GeneRegionTrack objects. This includes the “rstarts” and “rends”. You can consult the help for GeneRegionTrack() (?GeneRegionTrack to see from in R) to see full parameters required for UcscTrack objects.
##Bringing in External data. Tracks from UCSC
Now we can compare the Ensembl gene builds from the two different sources.
Notable differences in the annotation include the absense of some transcipts due to the additional filter applied in our BiomartGeneRegionTrack object creation.
##Bringing in External data. Tracks from UCSC as DataTrack
By the same method we can take advantage of other types of UCSC data.
In this example we capture the Conservation in the phyloP100wayAll table over and around our previously investigated ChIP-seq reads peak.
Here we specify the data to be returned as a DataTrack object and the display type to be “hist”. Here we are creating a DataTrack so would consult DataTrack() help (?DataTrack) to get full parameter list.
conservationTrack <- UcscTrack(genome = "hg19", chromosome = "chr5",track = "Conservation", table = "phyloP100wayAll",from = 135313003, to = 135313570, trackType = "DataTrack",start = "start", end = "end", data = "score",type = "hist", window = "auto", col.histogram = "darkblue",fill.histogram = "darkblue", ylim = c(-3.7, 4),name = "Conservation") ##Bringing in External data. Tracks from UCSC as DataTrack
conservationTrack <- UcscTrack(genome = "hg19",
chromosome = "chr5",
track = "Conservation",
table = "phyloP100wayAll",
from = 135313003,to = 135313570,
trackType = "DataTrack",
start = "start", end = "end",
data = "score",type = "hist", window = "auto",
col.histogram = "darkblue",fill.histogram = "darkblue",
ylim = c(-3.7, 4),name = "Conservation") ##Bringing in External data. Tracks from UCSC as DataTrack
With the inclusion of conservation alongside the coverage from CTCF peaks we can see a spike in conservation around the CTCF peak summit. We include a relative scale and increase the size of text for completeness.
plotTracks(c(conservationTrack,peakReads,genomeAxis),
from=135313003,
to=135313570,
chromosome = "chr5",
type = c("hist","coverage"),
sizes = c(1,1,0.2),
cex=2) 
##Exercises
Time for exercises! Link here
##Solutions
Time for solutions! Link here